Annotation of src/usr.bin/mandoc/man.c, Revision 1.109
1.109 ! schwarze 1: /* $OpenBSD: man.c,v 1.108 2015/04/19 14:25:05 schwarze Exp $ */
1.1 kristaps 2: /*
1.52 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.97 schwarze 4: * Copyright (c) 2013, 2014, 2015 Ingo Schwarze <schwarze@openbsd.org>
1.74 schwarze 5: * Copyright (c) 2011 Joerg Sonnenberger <joerg@netbsd.org>
1.1 kristaps 6: *
7: * Permission to use, copy, modify, and distribute this software for any
1.2 schwarze 8: * purpose with or without fee is hereby granted, provided that the above
9: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 10: *
1.98 schwarze 11: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.2 schwarze 12: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.98 schwarze 13: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.2 schwarze 14: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
15: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
16: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
17: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 18: */
1.14 schwarze 19: #include <sys/types.h>
20:
1.1 kristaps 21: #include <assert.h>
1.76 schwarze 22: #include <ctype.h>
1.1 kristaps 23: #include <stdarg.h>
24: #include <stdlib.h>
25: #include <stdio.h>
26: #include <string.h>
27:
1.98 schwarze 28: #include "mandoc_aux.h"
29: #include "mandoc.h"
30: #include "roff.h"
1.59 schwarze 31: #include "man.h"
1.98 schwarze 32: #include "libmandoc.h"
1.106 schwarze 33: #include "roff_int.h"
1.1 kristaps 34: #include "libman.h"
35:
1.79 schwarze 36: const char *const __man_macronames[MAN_MAX] = {
1.3 schwarze 37: "br", "TH", "SH", "SS",
1.79 schwarze 38: "TP", "LP", "PP", "P",
1.1 kristaps 39: "IP", "HP", "SM", "SB",
40: "BI", "IB", "BR", "RB",
41: "R", "B", "I", "IR",
1.96 schwarze 42: "RI", "sp", "nf",
1.50 schwarze 43: "fi", "RE", "RS", "DT",
44: "UC", "PD", "AT", "in",
1.70 schwarze 45: "ft", "OP", "EX", "EE",
1.78 schwarze 46: "UR", "UE", "ll"
1.1 kristaps 47: };
48:
49: const char * const *man_macronames = __man_macronames;
50:
1.101 schwarze 51: static void man_descope(struct roff_man *, int, int);
52: static int man_ptext(struct roff_man *, int, char *, int);
53: static int man_pmacro(struct roff_man *, int, char *, int);
1.1 kristaps 54:
55:
1.16 schwarze 56: void
1.101 schwarze 57: man_endparse(struct roff_man *man)
1.1 kristaps 58: {
59:
1.92 schwarze 60: man_macroend(man);
1.1 kristaps 61: }
62:
63: int
1.101 schwarze 64: man_parseln(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 65: {
1.25 schwarze 66:
1.98 schwarze 67: if (man->last->type != ROFFT_EQN || ln > man->last->line)
1.89 schwarze 68: man->flags |= MAN_NEWLINE;
1.54 schwarze 69:
1.69 schwarze 70: return (roff_getcontrol(man->roff, buf, &offs) ?
1.79 schwarze 71: man_pmacro(man, ln, buf, offs) :
72: man_ptext(man, ln, buf, offs));
1.1 kristaps 73: }
74:
1.92 schwarze 75: static void
1.101 schwarze 76: man_descope(struct roff_man *man, int line, int offs)
1.52 schwarze 77: {
78: /*
79: * Co-ordinate what happens with having a next-line scope open:
80: * first close out the element scope (if applicable), then close
81: * out the block scope (also if applicable).
82: */
83:
1.92 schwarze 84: if (man->flags & MAN_ELINE) {
1.69 schwarze 85: man->flags &= ~MAN_ELINE;
1.92 schwarze 86: man_unscope(man, man->last->parent);
1.52 schwarze 87: }
1.92 schwarze 88: if ( ! (man->flags & MAN_BLINE))
89: return;
1.69 schwarze 90: man->flags &= ~MAN_BLINE;
1.92 schwarze 91: man_unscope(man, man->last->parent);
1.106 schwarze 92: roff_body_alloc(man, line, offs, man->last->tok);
1.52 schwarze 93: }
94:
1.1 kristaps 95: static int
1.101 schwarze 96: man_ptext(struct roff_man *man, int line, char *buf, int offs)
1.1 kristaps 97: {
1.27 schwarze 98: int i;
1.26 schwarze 99:
1.10 schwarze 100: /* Literal free-form text whitespace is preserved. */
101:
1.92 schwarze 102: if (man->flags & MAN_LITERAL) {
1.108 schwarze 103: roff_word_alloc(man, line, offs, buf + offs);
1.92 schwarze 104: man_descope(man, line, offs);
105: return(1);
1.10 schwarze 106: }
107:
1.92 schwarze 108: for (i = offs; buf[i] == ' '; i++)
1.10 schwarze 109: /* Skip leading whitespace. */ ;
1.18 schwarze 110:
1.71 schwarze 111: /*
112: * Blank lines are ignored right after headings
113: * but add a single vertical space elsewhere.
114: */
115:
1.92 schwarze 116: if (buf[i] == '\0') {
1.27 schwarze 117: /* Allocate a blank entry. */
1.92 schwarze 118: if (man->last->tok != MAN_SH &&
119: man->last->tok != MAN_SS) {
1.109 ! schwarze 120: roff_elem_alloc(man, line, offs, MAN_sp);
1.101 schwarze 121: man->next = ROFF_NEXT_SIBLING;
1.71 schwarze 122: }
1.68 schwarze 123: return(1);
1.10 schwarze 124: }
1.1 kristaps 125:
1.79 schwarze 126: /*
1.27 schwarze 127: * Warn if the last un-escaped character is whitespace. Then
1.79 schwarze 128: * strip away the remaining spaces (tabs stay!).
1.27 schwarze 129: */
1.18 schwarze 130:
1.27 schwarze 131: i = (int)strlen(buf);
132: assert(i);
1.18 schwarze 133:
1.27 schwarze 134: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
135: if (i > 1 && '\\' != buf[i - 2])
1.81 schwarze 136: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
137: line, i - 1, NULL);
1.18 schwarze 138:
1.27 schwarze 139: for (--i; i && ' ' == buf[i]; i--)
140: /* Spin back to non-space. */ ;
1.10 schwarze 141:
1.27 schwarze 142: /* Jump ahead of escaped whitespace. */
143: i += '\\' == buf[i] ? 2 : 1;
1.18 schwarze 144:
1.27 schwarze 145: buf[i] = '\0';
1.10 schwarze 146: }
1.108 schwarze 147: roff_word_alloc(man, line, offs, buf + offs);
1.28 schwarze 148:
149: /*
150: * End-of-sentence check. If the last character is an unescaped
151: * EOS character, then flag the node as being the end of a
152: * sentence. The front-end will know how to interpret this.
153: */
154:
155: assert(i);
1.72 schwarze 156: if (mandoc_eos(buf, (size_t)i))
1.69 schwarze 157: man->last->flags |= MAN_EOS;
1.10 schwarze 158:
1.92 schwarze 159: man_descope(man, line, offs);
160: return(1);
1.1 kristaps 161: }
162:
1.53 schwarze 163: static int
1.101 schwarze 164: man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 165: {
1.99 schwarze 166: struct roff_node *n;
1.91 schwarze 167: const char *cp;
1.99 schwarze 168: int tok;
1.84 schwarze 169: int i, ppos;
170: int bline;
1.91 schwarze 171: char mac[5];
1.1 kristaps 172:
1.59 schwarze 173: ppos = offs;
1.1 kristaps 174:
1.23 schwarze 175: /*
1.59 schwarze 176: * Copy the first word into a nil-terminated buffer.
1.91 schwarze 177: * Stop when a space, tab, escape, or eoln is encountered.
1.23 schwarze 178: */
1.27 schwarze 179:
1.59 schwarze 180: i = 0;
1.91 schwarze 181: while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
1.59 schwarze 182: mac[i++] = buf[offs++];
1.1 kristaps 183:
1.59 schwarze 184: mac[i] = '\0';
1.1 kristaps 185:
1.107 schwarze 186: tok = (i > 0 && i < 4) ? man_hash_find(mac) : TOKEN_NONE;
1.1 kristaps 187:
1.107 schwarze 188: if (tok == TOKEN_NONE) {
1.85 schwarze 189: mandoc_msg(MANDOCERR_MACRO, man->parse,
190: ln, ppos, buf + ppos - 1);
1.1 kristaps 191: return(1);
192: }
193:
1.91 schwarze 194: /* Skip a leading escape sequence or tab. */
195:
196: switch (buf[offs]) {
197: case '\\':
198: cp = buf + offs + 1;
199: mandoc_escape(&cp, NULL, NULL);
200: offs = cp - buf;
201: break;
202: case '\t':
203: offs++;
204: break;
205: default:
206: break;
207: }
208:
209: /* Jump to the next non-whitespace word. */
1.1 kristaps 210:
1.92 schwarze 211: while (buf[offs] && buf[offs] == ' ')
1.59 schwarze 212: offs++;
1.18 schwarze 213:
1.79 schwarze 214: /*
1.27 schwarze 215: * Trailing whitespace. Note that tabs are allowed to be passed
216: * into the parser as "text", so we only warn about spaces here.
217: */
1.18 schwarze 218:
1.92 schwarze 219: if (buf[offs] == '\0' && buf[offs - 1] == ' ')
1.81 schwarze 220: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
221: ln, offs - 1, NULL);
1.1 kristaps 222:
1.79 schwarze 223: /*
1.97 schwarze 224: * Some macros break next-line scopes; otherwise, remember
225: * whether we are in next-line scope for a block head.
1.21 schwarze 226: */
227:
1.97 schwarze 228: man_breakscope(man, tok);
1.84 schwarze 229: bline = man->flags & MAN_BLINE;
1.24 schwarze 230:
231: /* Call to handler... */
1.1 kristaps 232:
1.22 schwarze 233: assert(man_macros[tok].fp);
1.92 schwarze 234: (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf);
1.73 schwarze 235:
236: /* In quick mode (for mandocdb), abort after the NAME section. */
237:
1.92 schwarze 238: if (man->quick && tok == MAN_SH) {
1.80 schwarze 239: n = man->last;
1.98 schwarze 240: if (n->type == ROFFT_BODY &&
1.80 schwarze 241: strcmp(n->prev->child->string, "NAME"))
242: return(2);
243: }
1.1 kristaps 244:
1.79 schwarze 245: /*
1.84 schwarze 246: * If we are in a next-line scope for a block head,
247: * close it out now and switch to the body,
248: * unless the next-line scope is allowed to continue.
1.21 schwarze 249: */
250:
1.84 schwarze 251: if ( ! bline || man->flags & MAN_ELINE ||
252: man_macros[tok].flags & MAN_NSCOPED)
1.21 schwarze 253: return(1);
1.1 kristaps 254:
1.92 schwarze 255: assert(man->flags & MAN_BLINE);
1.69 schwarze 256: man->flags &= ~MAN_BLINE;
1.1 kristaps 257:
1.92 schwarze 258: man_unscope(man, man->last->parent);
1.106 schwarze 259: roff_body_alloc(man, ln, ppos, man->last->tok);
1.92 schwarze 260: return(1);
1.97 schwarze 261: }
262:
263: void
1.101 schwarze 264: man_breakscope(struct roff_man *man, int tok)
1.97 schwarze 265: {
1.99 schwarze 266: struct roff_node *n;
1.97 schwarze 267:
268: /*
269: * An element next line scope is open,
270: * and the new macro is not allowed inside elements.
271: * Delete the element that is being broken.
272: */
273:
1.107 schwarze 274: if (man->flags & MAN_ELINE && (tok == TOKEN_NONE ||
1.97 schwarze 275: ! (man_macros[tok].flags & MAN_NSCOPED))) {
276: n = man->last;
1.98 schwarze 277: assert(n->type != ROFFT_TEXT);
1.97 schwarze 278: if (man_macros[n->tok].flags & MAN_NSCOPED)
279: n = n->parent;
280:
281: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
282: n->line, n->pos, "%s breaks %s",
1.107 schwarze 283: tok == TOKEN_NONE ? "TS" : man_macronames[tok],
1.97 schwarze 284: man_macronames[n->tok]);
285:
1.106 schwarze 286: roff_node_delete(man, n);
1.97 schwarze 287: man->flags &= ~MAN_ELINE;
288: }
289:
290: /*
291: * A block header next line scope is open,
292: * and the new macro is not allowed inside block headers.
293: * Delete the block that is being broken.
294: */
295:
1.107 schwarze 296: if (man->flags & MAN_BLINE && (tok == TOKEN_NONE ||
1.97 schwarze 297: man_macros[tok].flags & MAN_BSCOPE)) {
298: n = man->last;
1.98 schwarze 299: if (n->type == ROFFT_TEXT)
1.97 schwarze 300: n = n->parent;
301: if ( ! (man_macros[n->tok].flags & MAN_BSCOPE))
302: n = n->parent;
303:
1.98 schwarze 304: assert(n->type == ROFFT_HEAD);
1.97 schwarze 305: n = n->parent;
1.98 schwarze 306: assert(n->type == ROFFT_BLOCK);
1.97 schwarze 307: assert(man_macros[n->tok].flags & MAN_SCOPED);
308:
309: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
310: n->line, n->pos, "%s breaks %s",
1.107 schwarze 311: tok == TOKEN_NONE ? "TS" : man_macronames[tok],
1.97 schwarze 312: man_macronames[n->tok]);
313:
1.106 schwarze 314: roff_node_delete(man, n);
1.97 schwarze 315: man->flags &= ~MAN_BLINE;
316: }
1.62 schwarze 317: }
318:
319: const struct mparse *
1.101 schwarze 320: man_mparse(const struct roff_man *man)
1.62 schwarze 321: {
322:
1.69 schwarze 323: assert(man && man->parse);
324: return(man->parse);
1.76 schwarze 325: }
326:
327: void
1.99 schwarze 328: man_deroff(char **dest, const struct roff_node *n)
1.76 schwarze 329: {
330: char *cp;
331: size_t sz;
332:
1.98 schwarze 333: if (n->type != ROFFT_TEXT) {
1.76 schwarze 334: for (n = n->child; n; n = n->next)
335: man_deroff(dest, n);
336: return;
337: }
338:
1.77 schwarze 339: /* Skip leading whitespace and escape sequences. */
1.76 schwarze 340:
1.77 schwarze 341: cp = n->string;
342: while ('\0' != *cp) {
343: if ('\\' == *cp) {
344: cp++;
345: mandoc_escape((const char **)&cp, NULL, NULL);
346: } else if (isspace((unsigned char)*cp))
347: cp++;
348: else
1.76 schwarze 349: break;
1.77 schwarze 350: }
1.76 schwarze 351:
352: /* Skip trailing whitespace. */
353:
354: for (sz = strlen(cp); sz; sz--)
355: if (0 == isspace((unsigned char)cp[sz-1]))
356: break;
357:
358: /* Skip empty strings. */
359:
360: if (0 == sz)
361: return;
362:
363: if (NULL == *dest) {
364: *dest = mandoc_strndup(cp, sz);
365: return;
366: }
367:
368: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
369: free(*dest);
370: *dest = cp;
1.4 schwarze 371: }